there are different people from different background working on cognitive sciences
They study some kind of representational structures..
Computer as metaphor for the brain
input---output
What is representation?
Representations are symbolic for a concept → surrogate that refers to a referent
Mental representations---- has certain meaning (Semantic)
Relation between a mental representation and its referent is usually intentional
→ some similarity between representation and its referent
→ aspects can be mapped onto each other
→ referent triggers the mental representation which triggers behavior related to referent
Many different types of representations:
● Digital → discrete set of values e.g. language → syntax
● Analog → continuous representation of information e.g. vision → visual imagery
● Propositions → relation between concepts talking ([Silvy], [CSAI students])
relationship (subject, object)
Neurons in rats’ brains were recorded when rats navigated their environment
John O’keefe / Edvard Moser / May-Britt Moser
Nobel winning experiment
spatial awareness
Place cells - Hippocampus
Grid cells---mental representation of neurons that contain overview of environment rather
than only current environment → help you navigate your environment
Example 2: mental representations in the human brain of image categories?
Brain regions
Mental representation of faces vs mental representation of scenes
Faces vs Scenes---- MRI Scans
Specialized brain area for representing faces: Fusiform Face Area i.e. Fusiform Face Area
produces output when someone perceives a face
,some basic representation seems to be similar across people but every human being is
differnet
Haxby et al., Annu Rev Neurosci., 2014
Facial pattern across regions-
Mental representations of such basic categories (faces, scenes) are very robust, even in
noisy human fMRI data
Even in real-time, at single-subject and single-MRimage level and with only an indirect cue
of the category on the screen
History of Cognitive Science
In the 19th ct., psychology separated from philosophy and became an experimental science
● 1920s - JB Watson, BF Skinner - BEHAVIORISM
Data science approaches in US-- focus on patterns--not looking for explanations
● 1950s - restoration of the mind at the center of the study
● The ‘cognitive revolution’ was greatly influenced by work in computer science and
theoretical linguistics
● B.F. Skinner, Verbal Behavior (1957) -
a behavioral treatment of language and communication; language learning by
association stimulus-reward
● Poverty-of-stimulus argument by Noam Chomsky (1959) considered a conclusive criticism
against behaviorism and, by some, the foundational text of cognitive psychology
Skinner: “Chomsky simply does not understand what I am talking about and I see no
reason to listen to him” (as cited by Andresen, 1991:57)
Chomsky: “[After the war, there was this idea of] we’ll do it the American way, not that old
European way. In this context, radical behaviorism just fits easily. In fact, the study of human
affairs was called behavioral science. It was a very strange notion. Behavior is evidence. It’s
not what you are studying; what you are studying is competence, capacity. But the study of
behavior is like calling physics “meter-readings science” because meter readings are the
data.” (in Virués-Ortega, The Case against BF Skinner 45 years later (2006))
Language acquisition device---
, Chomsky on cognitive Sciences
“Obviously, every speaker of a language has mastered and internalized a generative
grammar that expresses his knowledge of his language. This is not to say that he is aware of
the rules of the grammar or even that he can become aware of them.” (Aspects of the
Theory of Syntax, 1965)
● Cognitive states thus not necessarily accessible to consciousness
● We don’t have any conscious access to it
There are sentences gramattically correct but they actually don’t make any sense
Logic theorists
thinking machine
Victor Yngve--machine translation system
one of the first syntactic approaches
George Miller-magic number 7
What’s computation
The mind performs computations on representations
● Think of mathematical or syntactic operations, but also sensation, perception,
attention, memory, language in general, logical reasoning, decision making,
and problem solving
● General processes such as scanning, matching, sorting, and retrieving
Marr’s Three Levels
● David Marr’s research on computational approaches to various brain areas
- Three-level hypothesis for reasoning about complex systems (brain, computer, or
human behavior)
1. COMPUTATIONAL - generic specification of the problem we are faced with
(e.g., learning a function, or estimating uncertainty)
2. ALGORITHMIC - how the problem can be solved (e.g., through Bayesian
statistics or MachineLearning)
3. IMPLEMENTATIONAL - the “hardware” (e.g., neurons and synapses, or
transistors